strisunshinewentingwang/pytorch

Pull Requests Commits

Properly avoid wrapping numbers as tensors before backend (#96193)

ezyang committed 1 year ago

02a18b1a

[MPS] Fix in-place add and sub with alpha == 0.0 (#96184)

qqaatw committed 1 year ago

2f66b57a

Profile dynamo guards (#96119)

wconstab committed 1 year ago

d4f5f9fd

[TEST] Turn on unspecialize int dynamic training inductor CI (#96058)

ezyang committed 1 year ago

d0641ed2

inductor: align baddbmm behavior with eager mode for beta=0 and input has nan value (#96087)

XiaobingSuper committed 1 year ago

9a575e77

fix issue of baddbmm when out has nan value for beta=0 (#96086)

XiaobingSuper committed 1 year ago

ac77883e

Improve ASAN and TSAN handling in cmake (#93147)

cyyever committed 1 year ago

666efd8d

Correctly pass $@ to the runner benchmark script (#96190)

ezyang committed 1 year ago

8dbc5495

[Easy] Fix typo "steams" -> "streams" (#95706)

awgu committed 1 year ago

40ca20bb

[FSDP][Docs] Per-device NCCL stream is per PG (#95705)

awgu committed 1 year ago

803e3044

COO intersection primitives: performance improvement (#96094)

nikitaved committed 1 year ago

98a4d74a

Use maxint to bound integers. (#96121)

ezyang committed 1 year ago

98ff841a

Turn on unspecialize int dynamic inductor CI (#96034)

ezyang committed 1 year ago

a6e3e790

Add a sample for index_fill to test framework (#91534)

qqaatw committed 1 year ago

3326c14e

[Dynamo] No graph break on namedtuple and potential other functions (#96122)

yanboliang committed 1 year ago

12ab4f08

Dynamo.export to preserve names of args & kwargs (#95851)

BowenBao committed 1 year ago

8ca3c881

Refactor unittest around dynamo.export wrt function signature (#95850)

BowenBao committed 1 year ago

e38f48ff

Type annotate `dynamo.export` (#95742)

BowenBao committed 1 year ago

c5965042

Restore _graph_executor_optimize flag after JIT test_profiler (#96135)

huydhn committed 1 year ago

18e8aa95

[MPS] Add type promotion to `torch.addcmul` (#96164)

malfet committed 1 year ago

769cc8a6

Add Generator register for the privateuse1 backend (#93920)

heidongxianhua committed 1 year ago

7038458c

fix typo under aten/src/ATen/cudnn (#96063)

kiszk committed 1 year ago

e9ca9028

Upload torch dynamo performance stats to S3 before Rockset (#96165)

huydhn committed 1 year ago

19ee61f7

fix typo in comments under torch/csrc/distributed (#96062)

kiszk committed 1 year ago

29739942

[inductor] Refactor IR printing (#96024)

jansel committed 1 year ago

fe4fec37

Stop printing data dependent variable stacks by default (#96120)

ezyang committed 1 year ago

4ab4d881

[BC] Allow only `bool` tensors as mask in `masked_select` (#96112)

malfet committed 1 year ago

1cd0929b

enable taskset core pinning in addition to numactl (#96011)

min-jean-cho committed 1 year ago

e70ea8d5

[inductor] improve cpp vec implementation of square (#96072)

Valentine233 committed 1 year ago

e168dbb9

Fix broken contribution stats upload job (#96003)

PaliC committed 1 year ago

bf01caf2

Newer Older